Compare AI models

Specifications
Parameters
16B7B
Context window
32k8k
Benchmarks
HumanEval
57.3%26.2%
MMLU
55.7%62.5%
GSM8K
72%36.5%
Overview
CompanyDeepSeekMistral
Release dateMay 16 2024Sep 27 2023
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek V2 Lite leads Mistral 7B on 2 of the 3 benchmarks they both report (HumanEval, MMLU, GSM8K). Mistral 7B shipped 232 days before DeepSeek V2 Lite, so benchmark comparisons should account for the intervening progress.

DeepSeek V2 Lite has 16B parameters, while Mistral 7B has 7B. Context windows are 32k (DeepSeek V2 Lite) vs 8k (Mistral 7B).

On HumanEval, DeepSeek V2 Lite leads at 57.3% vs Mistral 7B at 26.2%. On MMLU, Mistral 7B leads at 62.5% vs DeepSeek V2 Lite at 55.7%. On GSM8K, DeepSeek V2 Lite leads at 72% vs Mistral 7B at 36.5%.